[
https://issues.apache.org/jira/browse/FLINK-36128?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17876650#comment-17876650
]
Leonard Xu commented on FLINK-36128:
------------------------------------
Thanks [~xiqian_yu] for the reporting, I adjust the issue title to make the
blocker clear
> The in-flight data will meet uncoverable issue when restore from checkpoint
> once old schema has been flushed to downstream systems
> ----------------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-36128
> URL: https://issues.apache.org/jira/browse/FLINK-36128
> Project: Flink
> Issue Type: Bug
> Components: Flink CDC
> Affects Versions: cdc-3.1.1
> Reporter: yux
> Assignee: yux
> Priority: Blocker
> Labels: pull-request-available
> Fix For: cdc-3.2.0
>
>
> Currently, default schema evolution mode "EVOLVE" could not handle
> exceptions, and might not be able to restore from existing state correctly
> after failover. Before we can "manually trigger checkpoint" that was
> introduced in Flink 1.19, making "LENIENT" a default option might be more
> suitable.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
